Confocal Laser Scanning Microscopy (CLSM) of

The introduction of uorochromes to characterize bone material oered new possibilities to gain dynamical histomorphometric parameters as already established in a large number of lab-oratories for routine-analysis. Recently developed applications exceed these standard analysis methods thus allowing imaging of three dimensional structures or to make use of the uo-rescent signal to predene measurement regions for other methods. Modern CLSM systems provide great advantages in terms of image contrast and spatial resolution compared to

con-ventional uorescent microscopes. Contrary to concon-ventional microscopes, these devices use lasers with dierent wavelengths as light sources, and a confocal setup assures a dened mea-surement volume in x,y and z direction.

Figure 2.18: Illustration of the beam path in a confocal laser scanning microscope: (1) lter (2) objective, (3) out-of-focus layer, (4) in-focus layer, (5) beam splitter, (6) detector, (7) pinhole; Reprinted from Reference [117] with permission from Springer-Verlag GmbH.

Technical setup

Figure 2.18 illustrates the beam path in a CLSM. The beam-splitter (5) can be designed as a simple half reective mirror or a wavelength sensitive mirror optimizing illumination and detection properties. The objective lens focuses the beam on the sample. According to the confocal setup, the emitted signal is guided back to the beam splitter and further focused with a convex lens on a pinhole plane (7). The measurement setup is executed in a way, that the focal plane at the pinhole represents the desired focal plane on the sample (4). Thus, signals that originate from sample regions below or above (3) the focal plane are projected in front of or behind the pinhole plane. Consequently, the pinhole size denes the information volume and thus the origin of the orescent radiation that is further processed at the detector. In a CLSM usually one or more photomultipliers are used as detectors exhibiting an acceptable sensitivity (about40 % quantum eciency), high countrate processing capabilities and there-fore good signal to noise ration.[117]

Resolution

To nd the optimal compromise between high spatial resolution and signal to noise ratio, besides setting laser intensity and detector gain (voltage applied in the photomultiplier). The adjustment of the pinhole size is crucial as follows. The lateral resolution is dened as the smallest distance between two points, that still allows identifying them as separate objects.

Figure 2.19: Intensity

Lateral image blurring due to diraction can never be fully avoided. Thus the image of a perfect point on the sample corresponds to a disc with a certain radius at the pinhole plane, the so-called Airy disc, or more exactly the Airy pattern if further orders of intensity maxima are taken into account (Figure 2.19). The Airy disc contains about 97 % of the total light while the rst halo contains with1.7 %the majority of the remaining intensity. Consequently, regarding the resolution vs. intensity issue, for most cases it makes sense to set the pinhole in a way that it ts the Airy disk while higher order maxima are excluded in the beam path. As a matter of fact it is not possible to make statements on the shape of objects smaller than the Airy disk, and also the disc radius r determines the smallest distance for distinguishing two objects and thus dening the lateral resolution. Formula 2.5 is a result of an estimation for the Airy radius for confocal microscopy. λ depicts the uorescence wavelength. The opening angle φ and the refraction index of the lense n refer to the objective parameters and can be summarized to the numerical aperture (NA). Depending on the numerical aperture, with high quality optics about 20−25 % of the emitted photons can be collected.

rconf ocal = 0.4·λ·(sin(ϕ)·n) = 0.4· λ

N A (2.5)

The resolution can be somewhat increased by reducing the pinhole size to diameters lower than the Airy 1 value, but this is accompanied by a strong decrease in intensity. On the other hand, if there are only minor requirements for the lateral resolution also big pinholes might be sucient. For our device the pinhole sizes reach from 10µm − 600 µm.

Basics of uorescence

To nd the optimal measurement setup for CLSM measurements of samples labeled with uorochromes, some fundamental knowledge on the physics of the emission of uorescence radiation is helpful as detailed described in in the textbooks (e.g. [118] and [117]). A

uo-Figure 2.20: Jablonksi diagram il-lustrating the principle of uores-cence radiation generation. S₀, S₁ and S₀ label the main energy shells of the involved molecule. Reprinted from Reference [118] with permis-sion from Elsevier.

rochrome molecule has the ability to absorb light of a dened wavelength thus performing an electronic transition to an excited state. (S₀ →S₁). During a delay time (some picoseconds) non-radiative transitions might occur until the system reaches the lowest vibrational energy level of the excited electronic state (internal conversion). Subsequently, the molecule relaxes to the ground stateS₀, which is accompanied by the emission of a uorescence photon, featur-ing a wavelength larger than the excitement radiation as illustrated in the Jablonksi diagram in Figure 2.20. The energy of the uorescent radiation is below the exaltation energy (Stokes shift), thus making it possible to discriminate between reected or scattered primary photons and uorescence (Figure 2.21).

Applications

According to the molecule's electronic and vibrational states, the uorescence spectrum but also the excitation spectrum dier for dierent uorochromes¹ (See Chapters 3.4 and 4.2.2).

In a CLSM characteristic spectral properties of dierent dyes like Tetracycline (for humans) and Calcein or Alizarin (for animals) allow to separate the signal from the unspecic auto uorescence and further to map the location of the various uorochromes independently.

The mentioned substances refer to those, which are most commonly used to label mineralizing bone tissue. Of course there are manifold other uorochromes designed for various applications providing insight into biological systems in vitro and in vivo as described in the literature [117].

Recently, a method was established to visualize the osteocyte-lacuna-canaliculi network mak-ing use of the uorescent character of Rhodamine6G. Kerschnitzki et al. showed that a Rho-damine solution can be used to stain all inner and outer surfaces of the tissue, like cortical and trabecular surface and borders of the haversian channels, osteocyte lacunae, and canaliculi [85, 119]. Making use of the 3D imaging capabilities of the CLSM, a routine was developed

1 http://www.lifetechnologies.com/at/en/home/life-science/cell-analysis/labeling-chemistry/uorescence-spectraviewer.html

Figure 2.21: Excitation and emission spectrum of Tetramethylrhodamine isothiocyanate (TRITC) in methanol; Reprinted from Reference [118] with permission from Elsevier.

to produce volumetric images of the OLCN and further to use a skeletonization and quanti-cation routine to gain network density and other parameters [85].

Additionally, we found that Rhodamine has an unspecic but high anity to the organic ma-trix. This oers comprehensive information, when investigating mineralization defects, which include regions that exhibit no contrast to the embedding material in qBEI (Chapter 3.7).

Chapter 3

Material, Methods and Methodological Developments

3.1 Routine Sample Preparation

All specimens used in the present studies are undecalcied bone samples dissected from fe-murs of humans or mice, embedded in polymethylmethacrylate (PMMA) using an established protocol [14, 93, 120]. Murine distal femurs were xed in 70 %ethanol immediately after dis-section while human samples were frozen (≈ −20°C) for storage and put in 70 % ethanol before to sample preparation.

Prior to the embedding procedure, water was removed by a dehydration series of 70 % - 80

% - 95 % - 100 % ethanol and residual fat was removed by putting the dehydrated sample in acetone over night. If desired a Rhodamine6G (AppliChem, St.Louis, USA)-staining pro-cedure as described in Chapter 2.2.4 was performed at this point. Subsequently, the sample was placed in PMMA and the hardening process was initiated in an incubator during careful control of the temperature (for more details see Reference [90, 95, 93]).

The PMMA blocks were then trimmed and a low-speed diamond saw (Buehler Isomet, Lake Blu, Illinois) was used to cut the embedded bone samples in the desired orientation.

Murine femurs were cut either in longitudinal direction (Figure 3.1a) facilitating measure-ments at the cortical bone of the femoral midshaft (blue), the metaphyseal (red) and epi-physeal (orange) cancellous bone or in transversal direction at the diaphysis of the femur (perpendicular to the long axis of the bone). While longitudinal sections exhibit cancellous and cortical bone, growth plate and articular cartilage, transversal sections of the diaphysis gain access to the whole femur cross sections but include no trabecular bone structure (Figure 3.1b). Consequently, various cutting directions provide access to dierent histological regions.

Thus, to decide the cutting orientation, the addressed problem needs to be taken into account.

Figure 3.1: Example of longitudinal (a) and transversal (b) cutting direction of a mouse femur. Three compartments are considered for evaluation: diaphyseal cortical bone (blue), metaphyseal spongiosa (red), and epiphyseal spongiosa (orange)